Editor's Note: This story, originally published in the February 1992 issue of Scientific American, is being re-posted in light of Steven Chu's nomination as U.S. secretary of energy.
Before you turn another page of this magazine, consider your actions carefully. Every time you wish to grasp a page, you must place one finger above the paper and another below so that the distance between each finger and the paper is about equal to the diameter of an atom. At that point, the electrons at the surface of your fingers repel the electrons on either side of the page. This slight redistribution of charges produces an electric field that is strong enough to allow you to squeeze the page between your fingers. Remarkably, by applying electric forces at the atomic scale, you can hold onto objects that are, on the whole, electrically neutral.
In contrast, manipulating neutral objects that are atomic in size is a formidable technical challenge. Charged objects are much easier to control because electric and magnetic fields can exert much stronger forces over them. Indeed, for more than a century, scientists have applied electromagnetic forces to manipulate charged particles such as electrons and ions from a distance. But only in the past few years have researchers been able to move neutral particles at a distance with any facility. In particular, investigators have developed instruments that use lasers to trap and manipulate atoms and micron-size particles with astonishing control. These innovations have quickly led to a wide range of applications. My research group and others have cooled atoms to temperatures near absolute zero-conditions that allow us to examine quantum states of matter and unusual interactions between light and ultra-cold atoms. We have begun to develop atomic clocks and extremely sensitive accelerometers. Our techniques are being applied to handle such individual molecules as large polymers. In addition, we have devised an "optical tweezers" that uses laser beams to hold and move organelles inside of cells without puncturing the intervening membranes.
Almost a decade before scientists learned how to control neutral particles at a distance with laser light, they achieved some of the same tasks using magnetic fields. They applied fields to focus atoms in beams and trap them. After learning how to trap atoms with laser light, they turned to the vast arsenal of laser techniques to gain precise control over neutral particles. The first trap for neutral particles was developed by Wolfgang Paul of the University of Bonn. In 1978 he and his colleagues succeeded in trapping neutrons in a magnetic field. Seven years later, using the same basic principles, William D. Phillips and his colleagues at the National Bureau of Standards were able to trap atoms.
The magnetic trap can hold onto particles that have magnetic properties similar to those of a tiny bar magnet. To be more precise, the particle must carry a small magnetic dipole moment. If such a particle is placed in a magnetic field whose strength varies from region to region, it will move toward the weakest or strongest part of the field, depending on the particle's orientation [see illustration on next page]. Paul realized that it is possible to design a magnetic field with a local minimum in the field strength, and if the magnetic dipole is originally aligned to seek a position where the field is weakest, it will remain aligned in the "weak field-seeking" orientation [see "Cooling and Trapping Atoms," by W. D. Phillips and H. ]. Metcalf; SCIENTIFIC AMERICAN, March 1987].
Atoms can also be trapped by laser light. Light can exert forces on atoms and other neutral particles because it carries momentum. If an atom is bombarded with a beam of light of a particular frequency, it will continuously absorb and reemit photons, the quanta of light. As the atom absorbs photons, it will receive a barrage of momentum kicks in the direction that the light beam propagates. The kicks add up to produce a "scattering" force, which is proportional to the momentum of each photon and the number of photons that the atom scatters per second. Of course, for every photon the atom absorbs, it must emit one. But because the photons are released with no preferred direction, the changes in momentum caused by the emission average to zero. Absorption and emission have the net effect of pushing the atom in the direction that the light travels.
The magnitude of this scattering force is quite low. If an atom absorbs a single photon, its change in velocity is tiny compared with the average velocity of atoms in a gas at room temperature. (The change is on the order of one centimeter per second, the crawling speed of an ant, whereas an atom at room temperature moves at the speed of a supersonic jet.)
This scattering force was first detected in 1933 , when Otto R. Frisch used it to deflect a beam of sodium atoms. He prepared the atoms by vaporizing sodium in a container. To form the beam, he allowed the atoms to pass through a hole in the container and a series of slits. Once established, the beam was bombarded with light from a sodium lamp. Although, on average, each sodium atom absorbed only a single photon, Frisch was able to detect a slight deflection of the beam.
The scattering force that Frisch generated was far too weak to capture atoms. Decades later workers realized that the photon-scattering rate could be increased to more than 10 million photons per second, corresponding to a force 100,000 times greater than the pull of gravity by the earth. The first dramatic demonstration of the scattering force on atoms was made by two separate groups led by Phillips and John L. Hall at the National Bureau of Standards. In 1985 they stopped a beam of atoms and reduced the temperature of the atoms from roughly 3 00 kelvins (room temperature) to 0. 1 kelvin.
The power of the scattering force attainable with lasers gave researchers hope that they could not only stop atoms but trap them as well. But attempts to configure several laser beams so that they could collect and concentrate atoms in some region of space seemed doomed to failure. According to a principle known as the Optical Earnshaw Theorem, it is impossible to fashion a light trap out of any configuration of light beams if the scattering force is proportional to the light intensity. The problem is that the beams cannot be arranged to generate only inward directed forces. Any light that enters a trapping region must eventually escape and must therefore carry outward directed forces as well. Even if Luke Skywalker were a physicist, the (scattering) force would not always be with him.
Fortunately, an atomic trap can be based on another kind of force that light can exert on atoms. To understand this force, it is instructive to consider how small particles can be attracted to a positively charged object, such as a glass rod rubbed with cat's fur. The rod produces an electric field that polarizes the particle. Consequently, the average position of positive charges in the particle will be slightly farther away from the rod than the average position of the negative charges. This asymmetric distribution of charge is said to have a dipole moment. The attractive dipole force exerted by the electric field on the negative charges of the particle is stronger than the repulsive force on the positive charges. As a result, the particle is pulled toward the regions where the electric field is strongest. Notice that this force is analogous to the magnetic dipole force first used to trap neutrons and atoms. If the charge on the rod were negative, the electric field would induce a dipole moment of reversed polarity, and the particle would still be attracted to regions of high electric field.
Because of the dipole force, atoms can be trapped by an electric field that has a local maximum of some point in space. Could such fields be produced by some clever arrangement of electric charges? For any system of fixed charges, the answer is no. Yet an electric field with a local maximum can be achieved in a dynamic system. In particular, because light is made up of rapidly oscillating electric and magnetic fields, a focused laser beam can produce an alternating electric field with a local maximum. When the field interacts with an atom, it alters the distribution of electrons around the atom, thereby inducing an electric dipole moment. The atom will thus be attracted to the local maximum in the field, just as the charged particle was drawn toward the rod.
The fact that the electric field changes rapidly does not present a problem. As the field changes polarity, the dipole moment of the atom also switches around. As long as the field changes at a rate slower than the natural oscillation frequencies of the atom, the dipole moment remains aligned with the field. The atom therefore continues to move toward the local maximum. As a result, this dipole force can be used to confine atoms. In 1968 Vladilen S. Letokhov first proposed that atoms could be trapped in a light beam using the dipole force, and 10 years later Arthur Ashkin of AT&T Bell Laboratories suggested a more practical trap based on focused laser beams.
Although the dipole-force trap is elegant in conception, it had practical problems. To minimize the scattering force, the light must be tuned well below the frequency at which the atoms readily absorb photons. At those large de tunings , the trapping forces are so feeble that atoms as cold as 0.01 kelvin cannot be held in the trap. Even when colder atoms were placed in the trap, they would boil out of the trap in a matter of a few thousandths of a second as a result of the ever present photon scattering. In addition, the task of injecting atoms into the trap seemed daunting because the volume of the trap would only be 0.001 cubic millimeter. For these reasons, the challenges to optical trapping seemed formidable.
Then, in 1985 , a scheme for a workable optical trap became apparent after atoms were laser cooled in all dimensions and to much lower temperatures than the stopped atomic beams. The laser-cooling idea was first proposed in 1975 by Theodor Hansch and Arthur Schawlow of Stanford University. In the same year, a similar scheme for cooling trapped ions with lasers was proposed by David J. Wineland and Hans G. Dehmelt of the University of Washington. The researchers predicted that an atom could be cooled if it is irradiated from two sides by laser light at a frequency slightly lower than the frequency needed for maximum absorption. If the atom moves in a direction oppo- sing one of the light beams, the light, from the atom's perspective, increases in frequency. The light that has been shifted up in frequency is then likely to be absorbed by the atom. The light that the atom absorbs exerts a scattering force that slows the atom down.
How does the atom interact with the light traveling in the same direction? The atom is less likely to absorb the light because the light, again from the atom's perspective, has been shifted down in frequency. The net effect of both of the beams is that a scattering force is generated, opposing the motion of the atom. The beauty of this idea is that an atom mOving in the opposite direction will also experience a scattering force dragging it toward zero velocity. By surrounding the atom with three sets of counterpropagating beams along three mutually perpendicular axes, the atom can be cooled in all three dimensions.
In 1985 Ashkin, Leo Hollberg, John E. Bjorkholm, Alex Cable and I at AT&T Bell Labs were able to cool sodium atoms to 240 millionths of a kelvin. Because the light field acts as a viscous force, we dubbed the combination of laser beams used to create the drag force "optical molasses." Although not a trap, the atoms were confined in the viscous medium for periods as long as 0.5 second, until eventually they would leak out of the cooling beams.
Optical molasses enabled us to solve the three major problems that stood in the way of constructing a laser trap. First, by cooling the atoms to extremely low temperatures, we could reduce the random thermal motions of the atoms, making them easy to trap. Second, we could easily load the atoms into the trap. Simply by focusing the trapping beam in the center of the optical molasses, atoms would be snagged as they randomly wandered into the trapping beam. Third, by alternating between trapping and cooling light, we could reduce the heating effects of the trapping light. A year after we had perfected optical molasses, atoms could finally be trapped with light.
Even with the loading technique used in our first trap, an optical trap with a larger capture volume was desirable. A trap that could use the scattering force would need much less light intensity, which meant the constraints imposed by the Optical Earnshaw Theorem had to be circumvented. The important clue of how to design such a trap came from David E. Pritchard of the Massachusetts Institute of Technology and Carl E. Wieman of the University of Colorado and their colleagues. They pointed out that if magnetic or electric fields that varied over space were applied to atoms, the scattering force caused by the laser light would not necessarily be proportional to the light intensity.
This suggestion led Jean Dalibard of the Ecole Normale Superieure in Paris to propose a "magneto-optic" trap, which used a weak magnetic field and circularly polarized light. In 1987 Pritchard's group and my own at AT&T collaborated to construct such a trap. Three years later Wieman's team went on to show that this technique could be used to trap atoms in a glass cell, using inexpensive diode lasers. Their method eliminated the precooling procedures needed in our first trapping experiments. The fact that atoms could be trapped in a sealed cell also meant rare species of atoms, such as radioactive isotopes, could be optically manipulated. The magneto-optic trap has become the most widely used optical trap today.
Meanwhile researchers were making rapid progress in laser cooling. Phillips and his colleagues discovered that under certain conditions, optical molasses could be used to cool atoms to temperatures far below the lower limit predicted by the existing theory. This discovery prompted Dalibard and Claude Cohen-Tannoudji of the College de France and the Ecole Normale and my group at Stanford to construct a new theory of laser cooling based on a complex but beautiful interplay between the atoms and their interaction with the light fields. Currently atoms can be cooled to a temperature with an average velocity equal to three and a half photon recoils. For cesium atoms, it means a temperature lower than three microkelvins.
Going beyond optical molasses, Cohen- Tannoudji, Alain Aspect, Ennio Arimondo, Robin Kaiser and Nathalie Vansteenkiste, then all at the Ecole Normale, invented an ingenious scheme capable of cooling helium atoms below the recoil velocity of a single scattered photon. Helium atoms have been cooled to two microkelvins along one dimension, and work is under way to extend this technique to two and three dimensions. This cooling method captures an atom in a well-defined velocity state in much the same way atoms were trapped in space in our first optical trap. As the atom scatters photons, its velocity randomly changes. The French experiment establishes conditions that allow an atom to recoil and land in a particular quantum state, which is a combination of two states with two distinct velocities close to zero. Once in this state, the chance of scattering more photons is greatly reduced, meaning that additional photons cannot scatter and increase the velocity. If the atom does not happen to land in this quantum state, it continues to scatter photons and has more opportunities to seek out the desired low-velocity state. Thus, the atoms are cooled by letting them randomly walk into a "velocity trapped" quantum state.
Besides the cooling and trapping of atoms, investigators have demonstrated various atomic lenses, mirrors and diffraction gratings for manipulating atoms. They have also fashioned devices that have no counterpart in light optics. Researchers at Stanford and the University of Bonn have made "atomic funnels" that transform a collection of hot atoms into a well-controlled stream of cold atoms. The Stanford group has also made an "atomic trampoline" in which atoms bounce off a sheet of light extending out from a glass surface. With a curved glass surface, an atom trap based on gravity and light can be made.
Clearly, we have learned to push atoms around with amazing facility, but what do all these tricks enable us to do? With very cold atoms in vapor form, physicists are in a position to study how the atoms interact with one another at extremely low temperatures. According to quantum theory, an atom behaves like a wave whose length is equal to Planck's constant divided by the particle's momentum. As the atom is cooled, its momentum decreases, thereby increasing its wavelength. At sl\fficiently low temperatures, the average wavelength becomes comparable to the average distance between the atoms. At these low temperatures and high densities, quantum theory says that a significant fraction of all the atoms will condense into a single quantum ground state. This unusual form of matter, called a Bose-Einstein condensation, has been predicted but never observed in a vapor of atoms. Thomas]. Greytak and Daniel Kleppner of M.LT. and look T. M. Walraven of the University of Amsterdam are trying to achieve such a condensation with a collection of hydrogen atoms in a magnetic trap. Meanwhile other groups are attempting the same feat in a laser-cooled sample of alkali atoms such as cesium or lithium.
Atom-manipulation techniques are also offering new opportunities in highresolution spectroscopy. By combining several such techniques, the Stanford group has created a device that will allow the spectral features of atoms to be measured with exquisite accuracy. We have devised an atomic fountain that launches ultra-cold atoms upward gently enough to have gravity turn them around. Atoms for the fountain are collected by a magneto-optic trap for 0. 5 second. After that amount of time, about 10 million atoms are launched upward at a velocity of roughly two meters per second. At the top of the trajectory, an atom is probed with two pulses of microwave radiation separated in time. If the frequency of the radiation is properly tuned, the two pulses cause the atom to change from one quantum state to another. (Norman Ramsey shared the Nobel Prize in Physics in 1989 for inventing and applying this technique.) In our first experiment we measured the energy difference between two states of an atom with a resolution of two parts in 100 billion.
How does the fountain make such precise measurements possible? First, the atoms fall freely and are easy to shield from any perturbation that might alter their energy levels. Second, such measurements are limited in precision by the Heisenberg uncertainty prinCiple. This principle states that the resolution of an energy measurement will be limited to Planck's constant divided by the time of the "measurement." In our case, this time corresponds to the time between the two microwave pulses. With an atomic fountain the measurement time for unperturbed atoms can be as long as one second, a period impossible with atoms at room temperature.
Because the atomic fountain allows extremely precise measurements of the energy levels of atoms, it may be possible to adapt the device to make an improved atomic clock. At present, the world time standard is defined by the energy difference between two particular energy levels in ground states of the cesium atom. Two years after the first atomic fountain, the group at the Ecole Normale used a fountain to measure the "clock transition" in the cesium atom with high precision. These two experiments suggested that a properly engineered instrument might be able to measure the absolute frequency of this transition to one part in 10^16, 1,000 times better than the accuracy of our best clocks. Lured by this potential, more than eight groups around the world are now trying to improve the cesium time standard with an atomic fountain.
Another application being intensively studied is atom interferometry. The first atom interferometers were built in 1991 by investigators at the University of Konstanz, M.LT., the Physikalisch-Technische Bundesanstalt and Stanford.
An atom interferometer splits an atom into two waves separated in space. The two parts of the atom are then recombined and allowed to interfere with each other. The simplest example of such a splitting occurs when the atom is made to go through two separated mechanical slits. If the atom is recombined after passing through the slits, wavelike interference fringes can be observed. The interference effects from atoms dramatically demonstrate the fact that their behavior needs both a wave and a particle description.
More important, atom interferometers offer the possibility of measuring physical phenomena with high sensitivity. In the first demonstration of the potential sensitivity, Mark Kasevich and I have created an interferometer that uses slow atoms. The atoms were split apart and recombined in a fountain. With this instrument we have already shown that the acceleration of gravity can be measured with a resolution of at least three parts in 100 million, and we expect another 100-fold improvement shortly. Previously, the effects of gravity on an atom have been measured at a level of roughly one part in 100.
In recent years the work on atom trapping has stimulated renewed interest in manipulating other neutral particles. The basic principles of atom trapping can be applied to micron-size particles, such as polystyrene spheres. The intense electric field at the center of a focused laser beam polarizes the particle, just as it would polarize an atom. The particle, like an atom, will also absorb light of certain frequencies. Glass, for example, strongly absorbs ultraviolet radiation. But as long as the light is tuned below absorption frequency, the particle will be drawn into the region of highest laser intensity.
In 1986 Ashkin, Bjorkholm, ]. B. Dziedzic and I showed that particles that range in size between 0.02 and 10 microns can be trapped in a single focused laser beam. In 1970 Ashkin trapped micron-size latex spheres suspended in water in between two fo- cused, counterpropagating beams of light [see "The Pressure of Laser Light," by Arthur Ashkin; SCIENTIFIC AMERICAN, February 1972]. But only much later was it realized that if a single beam is focused tightly enough, the dipole force would suffice to overcome the scattering force that pushes the particle in the direction that the laser beam is traveling.
The great advantage of using a single beam is that it can be used as an optical tweezers to manipulate small particles. The optical tweezers can easily be integrated with a conventional microscope by introducing the laser light into the body of the scope and focusing it with the viewing objective. A sample placed on an ordinary microscope slide can be viewed and manipulated at the same time by moving the focused laser beam. One application of the optical tweezers, discovered by Dziedzic and Ashkin, has captured the imagination of biologists.
They found that the tweezers can handle live bacteria and other organisms without apparent damage. The ability to trap live organisms without harm is surprising, considering that the typical laser intensity at the focal point of the optical tweezers is about 10 million watts per square centimeter. It turns out that as long as the organism is very nearly transparent at the frequency of the trapping light, it can be cooled effectively by the surrounding water. To be sure, if the laser intensity is too high, the creature can be "optocuted."
Many applications have been found for the optical tweezers. Ashkin showed that objects within a living cell can be manipulated without puncturing the cell wall. Steven M. Block and his colleagues at the Rowland Institute in Cambridge, Mass., and at Harvard University have studied the mechanical properties of bacterial flagella. Michael W. Berns and his co-workers at the University of California at Irvine have manipulated chromosomes inside a cell nucleus.
Optical tweezers can be used to examine even smaller biological systems. My colleagues Robert Simmons, Jeff Finer, James A. Spudich and I are applying the optical tweezers to study muscle contraction at the molecular level. Related studies are being carried out by Block and also by Michael P. Sheetz of Duke University. One of the goals of this work is to measure the force generated by a single myosin molecule pulling against an actin filament. We are probing this "molecular motor" by attaching a polystyrene sphere to an actin filament and using the optical tweezers to grab onto the bead. When the myosin head strokes against the actin filament, the motion is sensed by a photodiode at the viewing end of the microscope. A feedback circuit then directs the optical tweezers to pull against the myosin in order to counteract any motion. In this way, we have measured the strength of the myosin pull under tension.
On an even smaller scale, Spudich, Steve Kron, Elizabeth Sunderman, Steve Quake and I are manipulating a single DNA molecule by attaching polystyrene spheres to the ends of a strand of DNA and holding the spheres with two optical tweezers. We can observe the molecule as we pull on it by staining the DNA vvith dye molecules, illuminating the dye with green light from an argon laser and detecting the fluorescence with a sensitive video camera. In our first experiments we measured the elastic properties of DNA. The two ends were pulled apart until the molecule was stretched out straight to its full length, and then one of the ends was released. By studying how the molecule springs back, we can test basic theories of polymer physics far from the equilibrium state.
The tweezers can also be used to prepare a single molecule for other ex- periments. By impaling the beads onto the microscope slide and increasing the laser power, we found that the bead can be "spot-welded" to the slide, leaving the DNA in a stretched state. That technique might be useful in preparing long strands of DNA for examination with state-of-the-art microscopes. Ultimately, we hope to use these manipulation abilities to examine the motion of enzymes along the DNA and to address questions related to gene expression and repair.
It has only been six years since workers have stopped atoms, captured them in optical molasses and made the first atom traps. Optical traps, to paraphrase a popular advertising slogan, have enabled us to "reach out and touch" particles in powerful new ways. We have shown that if we can "see" an atom or microscopic particle, we may be able to hold onto it regardless of intervening membranes. It has been a personal joy to see how esoteric conjectures in atomic physics have blossomed: the techniques and applications of laser cooling and trapping have gone well beyond our dreams during those early days. We now have important new tools for physics, chemistry and biology.